My name is Pierre, I'm currently employed as work experience at Clermont-Ferrand's university hospital (Centre Hospitalier Universitaire Gabriel Montpied).
The main goal was to develop a NGS pipeline to analyse bacterial genomes looking for sequence types, resistance and virulence genes and plasmids with several tools like trimmomatic/trimgalore; Srst2; MaximumCommonGenomePhylogeny and the best assembler I can find.
I managed to analyse and retrieve the sequence types of numerous strains of E. Coli, E. Cloacae and K. Pneumoniae and resistance genes as well.
Everything was doing fine but I recently ran into an issue, for the gene detection, on several sequenced and trimmed reads of genomes of the species mentioned above, I used the Virulence Factors DataBase hosted at http://www.mgc.ac.cn, well formatted with the scripts provided with Srst2 (as advised in the HowTo).
For most strains it is doing fine but for the rest it seems Srst2 is not able to provide a result, it displays a row of "?" instead of the alleles found or the "-" meaning no allele was found. For some strains relaunching the analysis made the issue vanish but I still get the issue with some data sets.
I double checked the dependencies, I'm using python 2.7.6, scipy 0.14.0, bowtie2 2.1.0 and finally samtools 0.1.18.
I thought this could come from the reads so I tried to trim with an other software but didn't see any significant difference.
I'm currently investigating the sequences by comparing assembled genomes for a dataset which worked with the ones which didn't, even if it's my best lead, I'm clueless.
Here's what the results looks like for Klebsiella strains with the Shigella virulence factors database.
(Files attached as well but I also have an issue where the columns are duplicated.)
The weird point is that strains pass the analysis on some databases but not all of them.
The command line was:
"srst2 --input_pe Seq/*.fq.gz --output genes_virulence --min_depth 10 --min_edge_depth 5 --min_coverage 90 --forward _R1_val_1 --reverse --R2_val_2 --gene_db VFDB/*_clustered.fasta --report_new_consensus"
The 'Seq' directory contains only the pair-end reads and the 'VFDB' directory also contains only the virulence factors databases.
I'll be truly thankful if you have some leads.
But I also would like to thank you guys for your time and I'm truly sorry to bother you.
Best regards,
Pierre C.